Thursday, February 13, 2025

Data acquisition

Cross-validation and model fitting

  1. Now that we have a draft approach for the tidymodels workflow (106; spatiotemporal_cv.R), implement it in targets on actual data.

Issue backlog

I think my major problem with issues is that they have been too large in the past. Change to more atomic issues so they can be checked off in an afternoon. This should help both keep the project moving and give me a sense of accomplishment at the end of the day. Working on large, nebulous issues is demoralizing.

  1. Started a new issue and branch, #107, to come up with quarto document for evaluating model performance and some light EDA.
  2. Created a quarto report that gets generated every time the modeling pipeline is run. The report should be named based on the date and include the code commit hash in the header.

General

   *** caught bus error ***
address 0x104dce5e8, cause 'invalid alignment'